- Email: [email protected]
Prediction of Preterm Birth: Nonsonographic Cervical Methods
Prediction of Preterm Birth: Nonsonographic Cervical Methods Michael G. Ross, MD, MPH,*,† and Marie H. Beall, MD*,† A short cervix in the second trimester is a powerful predictor of preterm birth risk. Multiple cervical length screens for patients in midpregnancy will likely become the standard of obstetrical care as a result of the development of effective methods (eg, cerclage, progesterone) to prevent early delivery in patients with a short cervix. Because of the high cost and infrastructure requirements, providing multiple cervical length evaluations through transvaginal ultrasound will likely be a significant barrier to universal screening. A cost-effective, low-technology method of cervical length screening is necessary to implement such programs. Available data suggest that digital examination is not sufficiently sensitive and reproducible to reliably screen for short cervix in presymptomatic patients in the mid trimester. New modalities for nonsonographic cervical length assessment (ie, Cervilenz) provide for a cost-effective, sensitive, and reproducible method of screening patients for short cervical length, which deserves further research in comparing its efficacy to sonographic cervical length. Semin Perinatol 33:312-316 © 2009 Elsevier Inc. All rights reserved. KEYWORDS cervical measurements, cervilenz, short cervix
T
he role of cervical length as measured by transvaginal ultrasound (TVU) in the prediction of preterm delivery is now well established. Whether a short cervix is associated with preterm delivery because of cervical insufficiency or preterm labor (PTL) or premature rupture of membranes, or both, remains uncertain, although there is likely a continuum between cervical insufficiency and PTL or premature rupture of membranes. With the advent of prophylactic measures for women with short cervix, there have been numerous proposals for the screening of both high-risk (eg, prior preterm delivery) and low-risk women with repeated cervical length assessments throughout midgestation. There is little data to suggest an optimal screening regimen, but the NIH MaternalFetal Network trial establishing the predictive value of TVU in high-risk patients screened patients with cervical length assessments every 2 weeks, from 16 to 22 weeks gestation.1 As noted by Spong,2 all pregnant women in the near future may undergo 5 or more cervical length assessments (from 16 *Department of Obstetrics and Gynecology, Harbor-UCLA Medical Center, Torrance, CA. †Department of Obstetrics and Gynecology, Geffen School of Medicine at UCLA, Los Angeles, CA. Dr Ross is a shareholder and Medical Director for Cervilenz. Address reprint requests to Michael G. Ross, MD, MPH, Department of Obstetrics and Gynecology, Harbor-UCLA Medical Center, 1000 W Carson St, Box 3, Torrance, CA 90502. E-mail: [email protected]
312
0146-0005/09/$-see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1053/j.semperi.2009.06.004
to 24 weeks gestation) to identify those who would benefit from cerclage or progesterone therapy. Additional cervical length assessments may be used for a follow-up of women with borderline cervical lengths or those receiving treatment. As discussed in this issue, in Prediction of Preterm Birth: Cervical Sonography by Mella and Berghella, TVU cervical length assessment serves as the gold standard and the preferred diagnostic test for a short cervix.3 The recommended technique for TVU4 in high-risk women includes a maximum of 6 cervical length measurements, with a recommended scan time of at least 5 minutes. Because considerable skill is required for accurate TVU cervical length measurement, it has also been suggested that providers be specifically certified to perform TVU cervical length. The cost of performing multiple additional ultrasound examinations by highly trained personnel in ⬎4 million pregnancies per year in the United States may well strain the resources available, both in health care financing as well as personnel and equipment. In conjunction with the standard 3 prenatal visits during this period (about 16, 20, 24 weeks), biweekly TVU screening by a maternal-fetal specialist would result in a maximum of 6-8 visits, creating a significant burden on the patient in regard to the time, travel, and logistics of medical appointments. In view of these challenges, it is important to develop alternative screening tests, with the goal of identifying a subset of women at highest risk for a short cervix, who would
Prediction of preterm birth then benefit from a diagnostic TVU cervical length assessment. From a patient’s perspective, it would be advantageous to perform screening cervical length assessments at the time and location of the routine prenatal visits at 16, 20, and 24 weeks; should they be indicated, additional screens performed at 18 and 22 weeks would add only 2 visits to the patient’s medical care regimen. The question remains whether a satisfactory screening examination is available. In this chapter, we will review nonsonographic alternatives to TVU for the assessment of cervical length, and will describe the data suggesting that an objective assessment of cervical portio length (CPL), performed at a routine prenatal care visit, may serve as a valuable screening test for patients at risk of preterm birth.
Digital Examination Digital estimation of CPL, performed as part of a bimanual examination, has been the most common means of evaluating cervical length at physical examination. The technique of digital cervical length estimation is not standardized, and to our knowledge there is no detailed description of the technique found in standard texts. Adding to the inaccuracy, cervical length is often expressed as percent effacement of the cervix, although the length of the “non-effaced” cervix depends on the individual and the time during pregnancy. Although some authorities have recommended using an estimate of the cervical length in centimeters,5 this approach has not been uniformly adopted. There are questions regarding the reliability and accuracy of digital examinations and the correlation of digital CPL with TVU measurements, perhaps, because of the lack of standardization. Goldberg et al6 compared independent, blinded digital cervical examinations by 2 examiners as well as independent, blinded TVU cervical length measurements, also by 2 examiners. These authors demonstrated that endovaginal ultrasonography was significantly more reproducible than digital examination, when agreement between examiners was defined as either ⫾4 or ⫾10 mm. Of note, TVU measurements averaged 20 mm greater than the digital measurements. The authors, however, do not report how the digital examination was performed (eg, where on the cervix the measurement was made, the degree of pressure on the vaginal fornix), an omission common to nearly all studies examining the accuracy of digital examination. Richey et al7 studied 100 women in the third trimester who presented with complaints of contractions or rupture of membranes. Cervical length was measured digitally and by transperineal ultrasound. There was a clinically and statistically significant correlation between the digital and transperineal measurements. In contrast to the study by Goldberg, digital examinations averaged only 2 mm less than transperineal measurements. Although extensive detail is provided concerning the technique for transperineal ultrasound, there is again little information regarding the specific technique for digital measurement. Sonek et al8 compared cervical length assessments by TVU with digital examination in 201 examinations of 83 gravidas throughout gestation. A total of 87% of cervical length measurements made by digital examination
313 were less than those made by TVU, with an average difference of 14 mm. In 13% of cases, the digital examination exceeded the TVU cervical length, with an average difference of 6 mm. The authors characterized the correlation between the 2 measurement techniques as “fair,” with a correlation coefficient of 0.49. Others have also reported a correlation between digital CPL and ultrasound measurements. Mahoney et al9 investigated the correlation of translabial ultrasound measurement of the cervix with digital CPL, expressed as percent effacement, in 109 third-trimester patients. A high degree of correlation was noted between ultrasound cervical length and percent effacement. Similarly, Iams et al4 reported a significant degree of correlation between TVU and Bishop’s score in the 3073 patients enrolled in the Maternal Fetal Medicine Unit Network trial of endovaginal sonography and preterm delivery. Finally, digital examination, abdominal ultrasound, and TVU were compared in 20 nongravid women undergoing hysterectomy.10 These authors suggested that digital examination underestimated cervical length by an average of 13.6 mm, whereas ultrasound measurements did not differ significantly from measurements made with a ruler after specimen removal. Once again, the technique for digital examination is not described, though the ultrasound and pathology techniques include extensive detail. Although these studies suggest, as expected, some correlation between the digital CPL and TVU cervical length, with the digital measurement measuring less, the technique of digital examination has not been standardized or studied with the same rigor as TVU, and the degree of correlation is moderate at best. In addition to the comparison of quantitative length measurements, several studies have been performed evaluating the predictive ability of digital vs ultrasound cervical length measurements for successful labor induction or risk for preterm delivery. In term patients, digital examination has been shown to be a useful clinical predictor. Chandra et al11 compared the 5 components of the Bishop’s score with TVU assessments for the prediction of successful labor induction in 120 patients at ⱖ41 weeks of gestation. No cervical ultrasound measurement was found to be significantly associated with successful induction or vaginal delivery within 24 hours. However, digital cervical effacement (ie, length) predicted both active labor in 12 hours and vaginal delivery in 24 hours. Reis et al12 also found the digital cervical examination of value for the prediction of vaginal delivery within 24 hours in women undergoing labor induction at term. In contrast, ultrasound cervical length was of marginal significance in predicting the outcome of induced labor. Rozenberg et al13 confirmed that the Bishop’s score was a better predictor of the time interval from induction to delivery and of successful vaginal delivery as compared with cervical length assessed by TVU. In contrast, Daskalakis et al14 performed TVU and Bishop’s score on 137 primiparous women undergoing labor induction. The Bishop’s score was not associated with mode of delivery, whereas TVU ⬍ 2.7 cm predicted vaginal delivery, with sensitivity and specificity each of about 75%. Gomez Laencina et al,15 Elghorori et al,16 and Uyar et al17 have also reported that TVU cervical length is better than the Bishop’s
M.G. Ross and M.H. Beall
314 score for predicting successful labor induction. Although Tan et al18 concluded that both TVU cervical length and Bishop’s scores are useful predictors, Tanir et al19 concluded both of these measures had limited predictive ability for successful labor induction. In summary, the published data is clearly divided on the usefulness of both digital CPL and TVU for the prediction of successful labor induction and delivery at term. Despite the potential value of digital cervical examination in term patients, digital CPL measurements have been less successful than TVU in predicting the risk of preterm delivery. With respect to screening for preterm delivery, Matijevic et al20 examined the predictive value of cervical length measurement by TVU and digital Bishop’s score in 282 low-risk women who were randomly assigned to a measurement technique in the second trimester. TVU cervical length was more sensitive and more specific than the Bishop’s score in predicting subsequent preterm delivery, although the sensitivities of the 2 techniques were only 57% and 33%, respectively. Zlatnik et al21 assessed 17 subjects from 16 to 26 weeks with TVU and digital examination. Only 3 patients met the study endpoint, which was the need for hospitalization before 26 weeks. The digital cervical length among these 3 patients was poorly predictive of preterm delivery. Notably, among 2 of the patients, digital cervical length actually increased 5-10 mm between biweekly examinations. McMahon et al22 studied 85 twin and 24 triplet pregnancies and found that all examinations tested performed better in predicting preterm delivery when performed at 24 than at 20 weeks. These authors did report that digital examination at 22 weeks had a 0% sensitivity for delivery before 32 weeks gestation, whereas TVU had a 27% sensitivity. Finally, Vayssiere et al23 studied 165 and 146 twin pregnancies at 21-23 and 26-28 weeks, respectively. They concluded that TVU predicted spontaneous delivery before 34 weeks better (as expressed as area under the receiver operator curve) than digital examination, when performed at a 27 week but not a 22 week visit. The positive predictive value of both techniques was 20% at 22 weeks; at 27 weeks the positive predictive values of TVU was 34%, but that of digital CPL was only 13%. These studies are not sufficiently large to make a final conclusion, but they certainly cast doubt on the usefulness of digital CPL to predict preterm delivery. Cervical length assessment has also been advocated as an aid in the diagnosis of PTL in symptomatic patients.24 Onderoglu25 assessed 90 patients admitted to the hospital with a diagnosis of PTL with intact membranes. Transperineal ultrasonography had a slightly superior receiver operator curve over percent effacement of the cervix for the prediction of which patient would deliver preterm. Gomez et al26 studied 59 patients admitted in PTL at 20-35 weeks. There was a significant relationship between preterm delivery and TVU cervical length, but not between delivery and percent effacement by digital examination. In contrast to these studies, Volumenie et al27 studied 59 patients admitted in PTL between 18 and 34 weeks of gestation and concluded that Bishop’s score by digital examination was as accurate as TVU in predicting preterm delivery among patients with idiopathic PTL. The combination of digital examination with ultra-
sonography for those at highest risk did not yield better results than digital examination alone. As with term patients, digital examination may have utility in predicting the risk of delivery in patients who are admitted in PTL. The combination of factors assessed by the Bishop’s score may be more predictive than cervical effacement alone. As noted earlier, there are no established criteria for the determination of digital CPL measurement in these studies; neither the degree of pressure to be placed on the vaginal fornix nor the location of the measurement (ie, anterior, posterior, or lateral) is described. Furthermore, the reproducibility of the measurements is questionable. Despite these issues, digital CPL, as a part of the Bishop’s score, has been successful in prediction of the outcome of labor in many studies. The difficulties in predicting preterm delivery by digital CPL may be due to the need for more accurate discrimination in presymptomatic patients, or it may be that longer cervices are inherently more difficult to assess digitally. For example, in Mahoney’s study,9 cervices rated uneffaced by digital examination demonstrated a wide range of TVU measured lengths (2.5-4.2 cm). The lack of consistency of digital CPL measurement makes it equally problematic as a screening test for patients who may benefit from a diagnostic TVU measurement. Although CPL measurement potentially has predictive value for both term and preterm delivery, a more consistent, reproducible measurement method is required.
Cervilenz Cervilenz (CerviLenz, Chagrin Falls, OH) is an FDA-approved, low-cost measuring device that allows a repeatable CPL measurement to be performed by a nurse or physician (Fig. 1). Cervilenz provides a low-cost screening tool for the identification of women who may benefit from a diagnostic TVU cervical length measurement, and possible cerclage or progesterone therapy. The Cervilenz measurement can be performed by nurses, nurse midwives, or physicians in the same setting in which prenatal care is provided, yielding an in-office, immediate CPL measure. Cervilenz measurements are performed under direct visualization after a speculum is inserted. The measuring probe is advanced along the outer aspect of the body of the cervix into the lateral fornix, until the examiner perceives gentle resistance of the vaginal apex. A plastic flange slides against the cervix, the Cervilenz is locked in place and the measurement read on the handle end of the device (Fig. 2). We have reported both the repeatability of Cervilenz measurements and the comparison of Cervilenz to TVU assessments among a population of women throughout the second
Figure 1 A schematic drawing of the Cervilenz device.
Prediction of preterm birth
315 suggest that digital assessment underestimates CPL, whereas the Cervilenz device permits a visualized and objective CPL measure in patients with PTL. Together, these results suggest that Cervilenz may provide a more reproducible and reliable screen for cervical length than digital examination. Both the digital examination and the Cervilenz device measure only the intravaginal portion of the cervix (CPL); the area of the internal os, where funneling occurs, is not examined. However, the degree and/or shape of internal os funneling is believed to have little predictive value beyond that of cervical length alone, as measured by TVU.30 More importantly, as demonstrated by To et al,31 funneling occurs in ⬍1% of patients with cervical length of ⬎30 mm. Thus, a Cervilenz measure of a normal CPL length would effectively rule out a short cervical length by TVU.
Conclusions
Figure 2 A schematic drawing, illustrating the proper positioning of the Cervilenz cervical flange against the cervical portion with the measuring rod along the lateral aspect of the body of the cervix.
and third trimesters. In a study performed in Southern California, a series of 189 women had cervical length measured by both Cervilenz and TVU, by examiners blinded to the alternative cervical length measurements.28 In the study, CPLs were measured at the left and right (3 and 9 o’clock), and anterior and posterior quadrants of the cervix (12 and 6 o’clock). The results demonstrated that left and right lateral CPL measures were highly correlated to each other (R ⫽ 0.95), whereas anterior lengths exceeded posterior lengths. Importantly, CPL measurements of ⬍30 mm were highly predictive of a short cervix by TVU, with a sensitivity of 88% and a specificity of 92%. Notably, Cervilenz is likely to cost a fraction of the $150-$200 expense for a TVU. Thus, Cervilenz represents a low cost, highly sensitive and specific screening tool that can potentially be used to identify patients who may subsequently benefit from TVU confirmation of short cervical length. Currently, a multicenter trial is in process to define the ROC curve for Cervilenz as a screening test to identify patients who require TVU at 18-24 weeks. Digital CPL being the technique currently available in prenatal offices, we sought to compare the accuracy of clinician digital examination with objective measurements using the Cervilenz device, in women presenting with symptoms of PTL.29 A total of 42 patients with singleton gestation from 24-34 weeks= gestation and cervical dilation ⬍ 3 cm underwent speculum examination and Cervilenz measurement. A second examiner, blinded to the Cervilenz results, digitally measured CPL. Digital CPL was significantly less than Cervilenz (2.88 vs 3.40 cm, P ⬍ 0.001), and in 36% of subjects this difference exceeded 1 cm. The discrepancy in CPL estimates persisted whether women were multiparous, had soft cervices, or had a history of preterm delivery. These results
In summary, obstetrical care is rapidly moving to incorporate multiple TVU cervical length measurements throughout the second and third trimesters, especially for women at high risk for preterm birth. In addition to its use in the prediction of PTL and its potential use in predicting the outcome of labor induction, cervical length has been suggested to predict the risk of bleeding in placenta previa32 and the likelihood of cesarean delivery at term.33 With the report of positive results from the NIH Maternal Fetal Medicine Network cerclage trial, the anticipated availability of progestational agents and recommendations from the ACOG regarding use of progesterone prophylaxis,34 there is a new awareness of the need for cervical length assessments. Much as the 1 hour glucose loading test serves as a screen to identify those women requiring the more expensive and time-consuming glucose tolerance test, so too an inexpensive, easily performed cervical length screen would conserve health care dollars by indicating which patients may benefit from a more costly, more specialized TVU diagnostic test. Although results of a multicenter trial are pending, the Cervilenz device has the potential to serve as this primary screening test.
References 1. Owen J, Yost N, Berghella V, et al: Mid-trimester endovaginal sonography in women at high risk for spontaneous preterm birth. J Am Med Assoc 286:1340-1348, 2001 2. Spong CY: Prediction and prevention of recurrent spontaneous preterm birth. Obstet Gynecol 110:405-415, 2007 3. Berghella V: Novel developments on cervical length screening and progesterone for preventing preterm birth. Br J Obstet Gynaecol 116:182187, 2009 4. Iams JD, Goldenberg RL, Meis PJ, et al: The length of the cervix and the risk of spontaneous premature delivery. National Institute of Child Health and Human Development Maternal Fetal Medicine Unit Network. N Engl J Med 334:567-572, 1996 5. Holcomb WL Jr, Smeltzer JS: Cervical effacement: variation in belief among clinicians. Obstet Gynecol 78:43-45, 1991 6. Goldberg J, Newman RB, Rust PF: Interobserver reliability of digital and endovaginal ultrasonographic cervical length measurements. Am J Obstet Gynecol 177:853-858, 1997 7. Richey SD, Ramin KD, Roberts SW, et al: The correlation between transperineal sonography and digital examination in the evaluation of the third-trimester cervix. Obstet Gynecol 85:745-748, 1995
316 8. Sonek JD, Iams JD, Blumenfeld M, et al: Measurement of cervical length in pregnancy: comparison between vaginal ultrasonography and digital examination. Obstet Gynecol 76:172-175, 1990 9. Mahony BS, Nyberg DA, Luthy DA, et al: Translabial ultrasound of the third-trimester uterine cervix. Correlation with digital examination. J Ultrasound Med 9:717-723, 1990 10. Jackson GM, Ludmir J, Bader TJ: The accuracy of digital examination and ultrasound in the evaluation of cervical length. Obstet Gynecol 79:214-218, 1992 11. Chandra S, Crane JM, Hutchens D, et al: Transvaginal ultrasound and digital examination in predicting successful labor induction. Obstet Gynecol 98:2-6, 2001 12. Reis FM, Gervasi MT, Florio P, et al: Prediction of successful induction of labor at term: role of clinical history, digital examination, ultrasound assessment of the cervix, and fetal fibronectin assay. Am J Obstet Gynecol 189:1361-1367, 2003 13. Rozenberg P, Chevret S, Chastang C, et al: Comparison of digital and ultrasonographic examination of the cervix in predicting time interval from induction to delivery in women with a low bishop score. Br J Obstet Gynaecol 112:192-196, 2005 14. Daskalakis G, Thomakos N, Hatziioannou L, et al: Sonographic cervical length measurement before labor induction in term nulliparous women. Fetal Diagn Ther 21:34-38, 2006 15. Gomez Laencina AM, Sanchez FG, Gimenez JH, et al: Comparison of ultrasonographic cervical length and the bishop score in predicting successful labor induction. Acta Obstet Gynecol Scand 86:799-804, 2007 16. Elghorori MR, Hassan I, Dartey W, et al: Comparison between subjective and objective assessments of the cervix before induction of labour. J Obstet Gynaecol 26:521-526, 2006 17. Uyar Y, Erbay G, Demir BC, et al: Comparison of the bishop score, body mass index and transvaginal cervical length in predicting the success of labor induction. Arch Gynecol Obstet 280:357-362, 2009 18. Tan PC, Vallikkannu N, Suguna S, et al: Transvaginal sonographic measurement of cervical length vs. bishop score in labor induction at term: tolerability and prediction of cesarean delivery. Ultrasound Obstet Gynecol 29:568-573, 2007 19. Tanir HM, Sener T, Yildiz Z: Digital and transvaginal ultrasound cervical assessment for prediction of successful labor induction. Int J Gynecol Obstet 100:52-55, 2008 20. Matijevic R, Grgic O, Vasilj O: Is sonographic assessment of cervical length better than digital examination in screening for preterm delivery in a low-risk population? Acta Obstet Gynecol Scand 85:1342-1347, 2006
M.G. Ross and M.H. Beall 21. Zlatnik FJ, Yankowitz J, Whitham J, et al: Vaginal ultrasound as an adjunct to cervical digital examination in women at risk of early delivery. Gynecol Obstet Invest 51:12-16, 2001 22. McMahon KS, Neerhof MG, Haney EI, et al: Prematurity in multiple gestations: identification of patients who are at low risk. Am J Obstet Gynecol 186:1137-1141, 2002 23. Vayssiere C, Favre R, Audibert F, et al: Cervical assessment at 22 and 27 weeks for the prediction of spontaneous birth before 34 weeks in twin pregnancies: is transvaginal sonography more accurate than digital examination? Ultrasound Obstet Gynecol 26:707-712, 2005 24. Alfirevic Z, len-Coward H, Molina F, et al: Targeted therapy for threatened preterm labor based on sonographic measurement of the cervical length: a randomized controlled trial. Ultrasound Obstet Gynecol 29: 47-50, 2007 25. Onderoglu LS: Digital examination and transperineal ultrasonographic measurement of cervical length to assess risk of preterm delivery. Int J Gynecol Obstet 59:223-228, 1997 26. Gomez R, Galasso M, Romero R, et al: Ultrasonographic examination of the uterine cervix is better than cervical digital examination as a predictor of the likelihood of premature delivery in patients with preterm labor and intact membranes. Am J Obstet Gynecol 171:956-964, 1994 27. Volumenie JL, Luton D, De Spirlet M, et al: Ultrasonographic cervical length measurement is not a better predictor of preterm delivery than digital examination in a population of patients with idiopathic preterm labor. Eur J Obstet Gynecol Reprod Biol 117:33-37, 2004 28. Ross MG, Cousins L, Baxter-Jones R, et al: Objective cervical portio length measurements: consistency and efficacy of screening for a short cervix. J Reprod Med 52:385-389, 2007 29. Burwick RM, Lee GT, Benedict JL, et al: Blinded comparison of cervical portio length measurements by digital examination vs Cervilenz. Am J Obstet Gynecol 200:e37-e39, 2009 30. Berghella V, Owen J, MacPherson C, et al: Natural history of cervical funneling in women at high risk for spontaneous preterm birth. Obstet Gynecol 109:863-869, 2007 31. To MS, Skentou C, Liao AW, et al: Cervical length and funneling at 23 weeks of gestation in the prediction of spontaneous early preterm delivery. Ultrasound Obstet Gynecol 18:200-203, 2001 32. Ghi T, Contro E, Martina T, et al: Cervical length and risk of antepartum bleeding in women with complete placenta previa. Ultrasound Obstet Gynecol 33:209-212, 2009 33. Smith GC, Celik E, To M, et al: Cervical length at mid-pregnancy and the risk of primary cesarean delivery. N Engl J Med 358:1346-1353, 2008 34. ACOG Committee. Opinion 419, 2008 (replaces no. 29 November 1, 2003). Use of progesterone to reduce preterm birth. Obstet Gynecol 112:963-965, 2008
T
he role of cervical length as measured by transvaginal ultrasound (TVU) in the prediction of preterm delivery is now well established. Whether a short cervix is associated with preterm delivery because of cervical insufficiency or preterm labor (PTL) or premature rupture of membranes, or both, remains uncertain, although there is likely a continuum between cervical insufficiency and PTL or premature rupture of membranes. With the advent of prophylactic measures for women with short cervix, there have been numerous proposals for the screening of both high-risk (eg, prior preterm delivery) and low-risk women with repeated cervical length assessments throughout midgestation. There is little data to suggest an optimal screening regimen, but the NIH MaternalFetal Network trial establishing the predictive value of TVU in high-risk patients screened patients with cervical length assessments every 2 weeks, from 16 to 22 weeks gestation.1 As noted by Spong,2 all pregnant women in the near future may undergo 5 or more cervical length assessments (from 16 *Department of Obstetrics and Gynecology, Harbor-UCLA Medical Center, Torrance, CA. †Department of Obstetrics and Gynecology, Geffen School of Medicine at UCLA, Los Angeles, CA. Dr Ross is a shareholder and Medical Director for Cervilenz. Address reprint requests to Michael G. Ross, MD, MPH, Department of Obstetrics and Gynecology, Harbor-UCLA Medical Center, 1000 W Carson St, Box 3, Torrance, CA 90502. E-mail: [email protected]
312
0146-0005/09/$-see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1053/j.semperi.2009.06.004
to 24 weeks gestation) to identify those who would benefit from cerclage or progesterone therapy. Additional cervical length assessments may be used for a follow-up of women with borderline cervical lengths or those receiving treatment. As discussed in this issue, in Prediction of Preterm Birth: Cervical Sonography by Mella and Berghella, TVU cervical length assessment serves as the gold standard and the preferred diagnostic test for a short cervix.3 The recommended technique for TVU4 in high-risk women includes a maximum of 6 cervical length measurements, with a recommended scan time of at least 5 minutes. Because considerable skill is required for accurate TVU cervical length measurement, it has also been suggested that providers be specifically certified to perform TVU cervical length. The cost of performing multiple additional ultrasound examinations by highly trained personnel in ⬎4 million pregnancies per year in the United States may well strain the resources available, both in health care financing as well as personnel and equipment. In conjunction with the standard 3 prenatal visits during this period (about 16, 20, 24 weeks), biweekly TVU screening by a maternal-fetal specialist would result in a maximum of 6-8 visits, creating a significant burden on the patient in regard to the time, travel, and logistics of medical appointments. In view of these challenges, it is important to develop alternative screening tests, with the goal of identifying a subset of women at highest risk for a short cervix, who would
Prediction of preterm birth then benefit from a diagnostic TVU cervical length assessment. From a patient’s perspective, it would be advantageous to perform screening cervical length assessments at the time and location of the routine prenatal visits at 16, 20, and 24 weeks; should they be indicated, additional screens performed at 18 and 22 weeks would add only 2 visits to the patient’s medical care regimen. The question remains whether a satisfactory screening examination is available. In this chapter, we will review nonsonographic alternatives to TVU for the assessment of cervical length, and will describe the data suggesting that an objective assessment of cervical portio length (CPL), performed at a routine prenatal care visit, may serve as a valuable screening test for patients at risk of preterm birth.
Digital Examination Digital estimation of CPL, performed as part of a bimanual examination, has been the most common means of evaluating cervical length at physical examination. The technique of digital cervical length estimation is not standardized, and to our knowledge there is no detailed description of the technique found in standard texts. Adding to the inaccuracy, cervical length is often expressed as percent effacement of the cervix, although the length of the “non-effaced” cervix depends on the individual and the time during pregnancy. Although some authorities have recommended using an estimate of the cervical length in centimeters,5 this approach has not been uniformly adopted. There are questions regarding the reliability and accuracy of digital examinations and the correlation of digital CPL with TVU measurements, perhaps, because of the lack of standardization. Goldberg et al6 compared independent, blinded digital cervical examinations by 2 examiners as well as independent, blinded TVU cervical length measurements, also by 2 examiners. These authors demonstrated that endovaginal ultrasonography was significantly more reproducible than digital examination, when agreement between examiners was defined as either ⫾4 or ⫾10 mm. Of note, TVU measurements averaged 20 mm greater than the digital measurements. The authors, however, do not report how the digital examination was performed (eg, where on the cervix the measurement was made, the degree of pressure on the vaginal fornix), an omission common to nearly all studies examining the accuracy of digital examination. Richey et al7 studied 100 women in the third trimester who presented with complaints of contractions or rupture of membranes. Cervical length was measured digitally and by transperineal ultrasound. There was a clinically and statistically significant correlation between the digital and transperineal measurements. In contrast to the study by Goldberg, digital examinations averaged only 2 mm less than transperineal measurements. Although extensive detail is provided concerning the technique for transperineal ultrasound, there is again little information regarding the specific technique for digital measurement. Sonek et al8 compared cervical length assessments by TVU with digital examination in 201 examinations of 83 gravidas throughout gestation. A total of 87% of cervical length measurements made by digital examination
313 were less than those made by TVU, with an average difference of 14 mm. In 13% of cases, the digital examination exceeded the TVU cervical length, with an average difference of 6 mm. The authors characterized the correlation between the 2 measurement techniques as “fair,” with a correlation coefficient of 0.49. Others have also reported a correlation between digital CPL and ultrasound measurements. Mahoney et al9 investigated the correlation of translabial ultrasound measurement of the cervix with digital CPL, expressed as percent effacement, in 109 third-trimester patients. A high degree of correlation was noted between ultrasound cervical length and percent effacement. Similarly, Iams et al4 reported a significant degree of correlation between TVU and Bishop’s score in the 3073 patients enrolled in the Maternal Fetal Medicine Unit Network trial of endovaginal sonography and preterm delivery. Finally, digital examination, abdominal ultrasound, and TVU were compared in 20 nongravid women undergoing hysterectomy.10 These authors suggested that digital examination underestimated cervical length by an average of 13.6 mm, whereas ultrasound measurements did not differ significantly from measurements made with a ruler after specimen removal. Once again, the technique for digital examination is not described, though the ultrasound and pathology techniques include extensive detail. Although these studies suggest, as expected, some correlation between the digital CPL and TVU cervical length, with the digital measurement measuring less, the technique of digital examination has not been standardized or studied with the same rigor as TVU, and the degree of correlation is moderate at best. In addition to the comparison of quantitative length measurements, several studies have been performed evaluating the predictive ability of digital vs ultrasound cervical length measurements for successful labor induction or risk for preterm delivery. In term patients, digital examination has been shown to be a useful clinical predictor. Chandra et al11 compared the 5 components of the Bishop’s score with TVU assessments for the prediction of successful labor induction in 120 patients at ⱖ41 weeks of gestation. No cervical ultrasound measurement was found to be significantly associated with successful induction or vaginal delivery within 24 hours. However, digital cervical effacement (ie, length) predicted both active labor in 12 hours and vaginal delivery in 24 hours. Reis et al12 also found the digital cervical examination of value for the prediction of vaginal delivery within 24 hours in women undergoing labor induction at term. In contrast, ultrasound cervical length was of marginal significance in predicting the outcome of induced labor. Rozenberg et al13 confirmed that the Bishop’s score was a better predictor of the time interval from induction to delivery and of successful vaginal delivery as compared with cervical length assessed by TVU. In contrast, Daskalakis et al14 performed TVU and Bishop’s score on 137 primiparous women undergoing labor induction. The Bishop’s score was not associated with mode of delivery, whereas TVU ⬍ 2.7 cm predicted vaginal delivery, with sensitivity and specificity each of about 75%. Gomez Laencina et al,15 Elghorori et al,16 and Uyar et al17 have also reported that TVU cervical length is better than the Bishop’s
M.G. Ross and M.H. Beall
314 score for predicting successful labor induction. Although Tan et al18 concluded that both TVU cervical length and Bishop’s scores are useful predictors, Tanir et al19 concluded both of these measures had limited predictive ability for successful labor induction. In summary, the published data is clearly divided on the usefulness of both digital CPL and TVU for the prediction of successful labor induction and delivery at term. Despite the potential value of digital cervical examination in term patients, digital CPL measurements have been less successful than TVU in predicting the risk of preterm delivery. With respect to screening for preterm delivery, Matijevic et al20 examined the predictive value of cervical length measurement by TVU and digital Bishop’s score in 282 low-risk women who were randomly assigned to a measurement technique in the second trimester. TVU cervical length was more sensitive and more specific than the Bishop’s score in predicting subsequent preterm delivery, although the sensitivities of the 2 techniques were only 57% and 33%, respectively. Zlatnik et al21 assessed 17 subjects from 16 to 26 weeks with TVU and digital examination. Only 3 patients met the study endpoint, which was the need for hospitalization before 26 weeks. The digital cervical length among these 3 patients was poorly predictive of preterm delivery. Notably, among 2 of the patients, digital cervical length actually increased 5-10 mm between biweekly examinations. McMahon et al22 studied 85 twin and 24 triplet pregnancies and found that all examinations tested performed better in predicting preterm delivery when performed at 24 than at 20 weeks. These authors did report that digital examination at 22 weeks had a 0% sensitivity for delivery before 32 weeks gestation, whereas TVU had a 27% sensitivity. Finally, Vayssiere et al23 studied 165 and 146 twin pregnancies at 21-23 and 26-28 weeks, respectively. They concluded that TVU predicted spontaneous delivery before 34 weeks better (as expressed as area under the receiver operator curve) than digital examination, when performed at a 27 week but not a 22 week visit. The positive predictive value of both techniques was 20% at 22 weeks; at 27 weeks the positive predictive values of TVU was 34%, but that of digital CPL was only 13%. These studies are not sufficiently large to make a final conclusion, but they certainly cast doubt on the usefulness of digital CPL to predict preterm delivery. Cervical length assessment has also been advocated as an aid in the diagnosis of PTL in symptomatic patients.24 Onderoglu25 assessed 90 patients admitted to the hospital with a diagnosis of PTL with intact membranes. Transperineal ultrasonography had a slightly superior receiver operator curve over percent effacement of the cervix for the prediction of which patient would deliver preterm. Gomez et al26 studied 59 patients admitted in PTL at 20-35 weeks. There was a significant relationship between preterm delivery and TVU cervical length, but not between delivery and percent effacement by digital examination. In contrast to these studies, Volumenie et al27 studied 59 patients admitted in PTL between 18 and 34 weeks of gestation and concluded that Bishop’s score by digital examination was as accurate as TVU in predicting preterm delivery among patients with idiopathic PTL. The combination of digital examination with ultra-
sonography for those at highest risk did not yield better results than digital examination alone. As with term patients, digital examination may have utility in predicting the risk of delivery in patients who are admitted in PTL. The combination of factors assessed by the Bishop’s score may be more predictive than cervical effacement alone. As noted earlier, there are no established criteria for the determination of digital CPL measurement in these studies; neither the degree of pressure to be placed on the vaginal fornix nor the location of the measurement (ie, anterior, posterior, or lateral) is described. Furthermore, the reproducibility of the measurements is questionable. Despite these issues, digital CPL, as a part of the Bishop’s score, has been successful in prediction of the outcome of labor in many studies. The difficulties in predicting preterm delivery by digital CPL may be due to the need for more accurate discrimination in presymptomatic patients, or it may be that longer cervices are inherently more difficult to assess digitally. For example, in Mahoney’s study,9 cervices rated uneffaced by digital examination demonstrated a wide range of TVU measured lengths (2.5-4.2 cm). The lack of consistency of digital CPL measurement makes it equally problematic as a screening test for patients who may benefit from a diagnostic TVU measurement. Although CPL measurement potentially has predictive value for both term and preterm delivery, a more consistent, reproducible measurement method is required.
Cervilenz Cervilenz (CerviLenz, Chagrin Falls, OH) is an FDA-approved, low-cost measuring device that allows a repeatable CPL measurement to be performed by a nurse or physician (Fig. 1). Cervilenz provides a low-cost screening tool for the identification of women who may benefit from a diagnostic TVU cervical length measurement, and possible cerclage or progesterone therapy. The Cervilenz measurement can be performed by nurses, nurse midwives, or physicians in the same setting in which prenatal care is provided, yielding an in-office, immediate CPL measure. Cervilenz measurements are performed under direct visualization after a speculum is inserted. The measuring probe is advanced along the outer aspect of the body of the cervix into the lateral fornix, until the examiner perceives gentle resistance of the vaginal apex. A plastic flange slides against the cervix, the Cervilenz is locked in place and the measurement read on the handle end of the device (Fig. 2). We have reported both the repeatability of Cervilenz measurements and the comparison of Cervilenz to TVU assessments among a population of women throughout the second
Figure 1 A schematic drawing of the Cervilenz device.
Prediction of preterm birth
315 suggest that digital assessment underestimates CPL, whereas the Cervilenz device permits a visualized and objective CPL measure in patients with PTL. Together, these results suggest that Cervilenz may provide a more reproducible and reliable screen for cervical length than digital examination. Both the digital examination and the Cervilenz device measure only the intravaginal portion of the cervix (CPL); the area of the internal os, where funneling occurs, is not examined. However, the degree and/or shape of internal os funneling is believed to have little predictive value beyond that of cervical length alone, as measured by TVU.30 More importantly, as demonstrated by To et al,31 funneling occurs in ⬍1% of patients with cervical length of ⬎30 mm. Thus, a Cervilenz measure of a normal CPL length would effectively rule out a short cervical length by TVU.
Conclusions
Figure 2 A schematic drawing, illustrating the proper positioning of the Cervilenz cervical flange against the cervical portion with the measuring rod along the lateral aspect of the body of the cervix.
and third trimesters. In a study performed in Southern California, a series of 189 women had cervical length measured by both Cervilenz and TVU, by examiners blinded to the alternative cervical length measurements.28 In the study, CPLs were measured at the left and right (3 and 9 o’clock), and anterior and posterior quadrants of the cervix (12 and 6 o’clock). The results demonstrated that left and right lateral CPL measures were highly correlated to each other (R ⫽ 0.95), whereas anterior lengths exceeded posterior lengths. Importantly, CPL measurements of ⬍30 mm were highly predictive of a short cervix by TVU, with a sensitivity of 88% and a specificity of 92%. Notably, Cervilenz is likely to cost a fraction of the $150-$200 expense for a TVU. Thus, Cervilenz represents a low cost, highly sensitive and specific screening tool that can potentially be used to identify patients who may subsequently benefit from TVU confirmation of short cervical length. Currently, a multicenter trial is in process to define the ROC curve for Cervilenz as a screening test to identify patients who require TVU at 18-24 weeks. Digital CPL being the technique currently available in prenatal offices, we sought to compare the accuracy of clinician digital examination with objective measurements using the Cervilenz device, in women presenting with symptoms of PTL.29 A total of 42 patients with singleton gestation from 24-34 weeks= gestation and cervical dilation ⬍ 3 cm underwent speculum examination and Cervilenz measurement. A second examiner, blinded to the Cervilenz results, digitally measured CPL. Digital CPL was significantly less than Cervilenz (2.88 vs 3.40 cm, P ⬍ 0.001), and in 36% of subjects this difference exceeded 1 cm. The discrepancy in CPL estimates persisted whether women were multiparous, had soft cervices, or had a history of preterm delivery. These results
In summary, obstetrical care is rapidly moving to incorporate multiple TVU cervical length measurements throughout the second and third trimesters, especially for women at high risk for preterm birth. In addition to its use in the prediction of PTL and its potential use in predicting the outcome of labor induction, cervical length has been suggested to predict the risk of bleeding in placenta previa32 and the likelihood of cesarean delivery at term.33 With the report of positive results from the NIH Maternal Fetal Medicine Network cerclage trial, the anticipated availability of progestational agents and recommendations from the ACOG regarding use of progesterone prophylaxis,34 there is a new awareness of the need for cervical length assessments. Much as the 1 hour glucose loading test serves as a screen to identify those women requiring the more expensive and time-consuming glucose tolerance test, so too an inexpensive, easily performed cervical length screen would conserve health care dollars by indicating which patients may benefit from a more costly, more specialized TVU diagnostic test. Although results of a multicenter trial are pending, the Cervilenz device has the potential to serve as this primary screening test.
References 1. Owen J, Yost N, Berghella V, et al: Mid-trimester endovaginal sonography in women at high risk for spontaneous preterm birth. J Am Med Assoc 286:1340-1348, 2001 2. Spong CY: Prediction and prevention of recurrent spontaneous preterm birth. Obstet Gynecol 110:405-415, 2007 3. Berghella V: Novel developments on cervical length screening and progesterone for preventing preterm birth. Br J Obstet Gynaecol 116:182187, 2009 4. Iams JD, Goldenberg RL, Meis PJ, et al: The length of the cervix and the risk of spontaneous premature delivery. National Institute of Child Health and Human Development Maternal Fetal Medicine Unit Network. N Engl J Med 334:567-572, 1996 5. Holcomb WL Jr, Smeltzer JS: Cervical effacement: variation in belief among clinicians. Obstet Gynecol 78:43-45, 1991 6. Goldberg J, Newman RB, Rust PF: Interobserver reliability of digital and endovaginal ultrasonographic cervical length measurements. Am J Obstet Gynecol 177:853-858, 1997 7. Richey SD, Ramin KD, Roberts SW, et al: The correlation between transperineal sonography and digital examination in the evaluation of the third-trimester cervix. Obstet Gynecol 85:745-748, 1995
316 8. Sonek JD, Iams JD, Blumenfeld M, et al: Measurement of cervical length in pregnancy: comparison between vaginal ultrasonography and digital examination. Obstet Gynecol 76:172-175, 1990 9. Mahony BS, Nyberg DA, Luthy DA, et al: Translabial ultrasound of the third-trimester uterine cervix. Correlation with digital examination. J Ultrasound Med 9:717-723, 1990 10. Jackson GM, Ludmir J, Bader TJ: The accuracy of digital examination and ultrasound in the evaluation of cervical length. Obstet Gynecol 79:214-218, 1992 11. Chandra S, Crane JM, Hutchens D, et al: Transvaginal ultrasound and digital examination in predicting successful labor induction. Obstet Gynecol 98:2-6, 2001 12. Reis FM, Gervasi MT, Florio P, et al: Prediction of successful induction of labor at term: role of clinical history, digital examination, ultrasound assessment of the cervix, and fetal fibronectin assay. Am J Obstet Gynecol 189:1361-1367, 2003 13. Rozenberg P, Chevret S, Chastang C, et al: Comparison of digital and ultrasonographic examination of the cervix in predicting time interval from induction to delivery in women with a low bishop score. Br J Obstet Gynaecol 112:192-196, 2005 14. Daskalakis G, Thomakos N, Hatziioannou L, et al: Sonographic cervical length measurement before labor induction in term nulliparous women. Fetal Diagn Ther 21:34-38, 2006 15. Gomez Laencina AM, Sanchez FG, Gimenez JH, et al: Comparison of ultrasonographic cervical length and the bishop score in predicting successful labor induction. Acta Obstet Gynecol Scand 86:799-804, 2007 16. Elghorori MR, Hassan I, Dartey W, et al: Comparison between subjective and objective assessments of the cervix before induction of labour. J Obstet Gynaecol 26:521-526, 2006 17. Uyar Y, Erbay G, Demir BC, et al: Comparison of the bishop score, body mass index and transvaginal cervical length in predicting the success of labor induction. Arch Gynecol Obstet 280:357-362, 2009 18. Tan PC, Vallikkannu N, Suguna S, et al: Transvaginal sonographic measurement of cervical length vs. bishop score in labor induction at term: tolerability and prediction of cesarean delivery. Ultrasound Obstet Gynecol 29:568-573, 2007 19. Tanir HM, Sener T, Yildiz Z: Digital and transvaginal ultrasound cervical assessment for prediction of successful labor induction. Int J Gynecol Obstet 100:52-55, 2008 20. Matijevic R, Grgic O, Vasilj O: Is sonographic assessment of cervical length better than digital examination in screening for preterm delivery in a low-risk population? Acta Obstet Gynecol Scand 85:1342-1347, 2006
M.G. Ross and M.H. Beall 21. Zlatnik FJ, Yankowitz J, Whitham J, et al: Vaginal ultrasound as an adjunct to cervical digital examination in women at risk of early delivery. Gynecol Obstet Invest 51:12-16, 2001 22. McMahon KS, Neerhof MG, Haney EI, et al: Prematurity in multiple gestations: identification of patients who are at low risk. Am J Obstet Gynecol 186:1137-1141, 2002 23. Vayssiere C, Favre R, Audibert F, et al: Cervical assessment at 22 and 27 weeks for the prediction of spontaneous birth before 34 weeks in twin pregnancies: is transvaginal sonography more accurate than digital examination? Ultrasound Obstet Gynecol 26:707-712, 2005 24. Alfirevic Z, len-Coward H, Molina F, et al: Targeted therapy for threatened preterm labor based on sonographic measurement of the cervical length: a randomized controlled trial. Ultrasound Obstet Gynecol 29: 47-50, 2007 25. Onderoglu LS: Digital examination and transperineal ultrasonographic measurement of cervical length to assess risk of preterm delivery. Int J Gynecol Obstet 59:223-228, 1997 26. Gomez R, Galasso M, Romero R, et al: Ultrasonographic examination of the uterine cervix is better than cervical digital examination as a predictor of the likelihood of premature delivery in patients with preterm labor and intact membranes. Am J Obstet Gynecol 171:956-964, 1994 27. Volumenie JL, Luton D, De Spirlet M, et al: Ultrasonographic cervical length measurement is not a better predictor of preterm delivery than digital examination in a population of patients with idiopathic preterm labor. Eur J Obstet Gynecol Reprod Biol 117:33-37, 2004 28. Ross MG, Cousins L, Baxter-Jones R, et al: Objective cervical portio length measurements: consistency and efficacy of screening for a short cervix. J Reprod Med 52:385-389, 2007 29. Burwick RM, Lee GT, Benedict JL, et al: Blinded comparison of cervical portio length measurements by digital examination vs Cervilenz. Am J Obstet Gynecol 200:e37-e39, 2009 30. Berghella V, Owen J, MacPherson C, et al: Natural history of cervical funneling in women at high risk for spontaneous preterm birth. Obstet Gynecol 109:863-869, 2007 31. To MS, Skentou C, Liao AW, et al: Cervical length and funneling at 23 weeks of gestation in the prediction of spontaneous early preterm delivery. Ultrasound Obstet Gynecol 18:200-203, 2001 32. Ghi T, Contro E, Martina T, et al: Cervical length and risk of antepartum bleeding in women with complete placenta previa. Ultrasound Obstet Gynecol 33:209-212, 2009 33. Smith GC, Celik E, To M, et al: Cervical length at mid-pregnancy and the risk of primary cesarean delivery. N Engl J Med 358:1346-1353, 2008 34. ACOG Committee. Opinion 419, 2008 (replaces no. 29 November 1, 2003). Use of progesterone to reduce preterm birth. Obstet Gynecol 112:963-965, 2008